This is LexFusion’s
annual Year in Review.
In January 2022,
we explored the culture of innovation (or lack thereof) in legal.
In January 2023,
we predicted Generative AI was posed to dominate the conversation (hard to recall now how controversial that was at the time) whether legal was ready or not (probably not).
Now, in 2024,
we are showing our work. We have published our 2500+ item, primary-source appendix for our presentations on GenAI and the economics of the legal market—the final touches on which delayed the publication of this year in review. Soon, we will make public an anonymized version of a red-team memo we wrote for a Global 50 law department.”
Our Year in Brief
2023 was another banner year for LexFusion.
First and foremost,
we lured the fabulous Christina Wojcik away from Citibank. Christina ran legal tech and innovation initiatives for Citi. At LexFusion, Christina is spearheading our relationships with in-house departments and focusing on transaction tech.
For LexFusion 2023 was filled with more than 2600 meetings, split almost evenly between law departments and law firms. Reflecting the reality that law departments of consequence are more dispersed than law firms of consequence, those meetings comprise 523 unique law departments and 230 unique law firms.
Our portfolio of member companies enjoyed similar success.
New Era ADR
secured a strategic investment from the Chicago Cubs* a few months after being named the official ADR platform of the United States Olympic & Paralympic Committee.*
Expert Institute
Ranked #1 in National Law Journal Best Of 2023 Awards*
Insight Optix
Released Evidence Optix® v3.0*
LexFusion also upped the ante on events.
In addition to our normal happy hours, dinners, and sporting events, we convened invite-only discussions under the Chatham House Rule in New York, San Francisco, and London to create a safe space for more than 100 law departments and 100 law firms to share their real-world experience to date deploying Generative AI.
Those GenAI roundtables merit reflection since we ended our 2022 year in review with both a prediction and a question predicated on our confidence that
GenAI would dominate the conversation in 2023:
In 2023, 90% of the population will believe AI is capable of generating, near instantly, a legal opinion or contract superior to the work product of 90% of junior lawyers.
What happens when the CFO hires the reinvigorated [Big Name Consultancy] Digital AI Transformation Team for a top-to-bottom efficiency review and, among many other recommendations with profound implications for the business, the resulting report plays to the CFO’s confirmation bias, finding that legal is one of many areas where low-level work can be expeditiously automated—to the point where, in many instances, legal can be bypassed entirely?
GenAI Roundtables
#realtalk.
We were so correct that GenAI would dominate the 2023 conversation that our prediction above seems almost trite in retrospect. Easy to forget how many people were dismissing ChatGPT as a novelty at this time last year. They had no idea what was coming. We, however, were trading on insider information.
Thus, we were in a privileged position to pose as prescient when we penned our GenAI primer PSA: ChatGPT is the trailer, not the movie. We had a front-row seat to law departments and law firms expending real resources to experiment with GPT-4 on real-world use cases.
We certainly contributed to this noise. Since December 2022, we’ve been presenting regularly on GenAI.
Indeed, we’ve gone through so many iterations of our GenAI presentation that we published our 2000+ item, primary-source, annotated appendix. While we frequently lead more mature discussions, the truth is most of our presentations have been 101 level introductions—in order to meet the majority of the audience where they are.
To advance the collective conversation, we ultimately decided to host our own events. Closed door. Invite only. No sponsors. Chatham House Rule.
The criteria for speaking: the organization had to (i) be far enough along in its deployment of GenAI on legal use cases to have real lessons learned and (ii) be willing to share their real-world experiences, positive and negative, with a degree specificity sufficient for those lessons learned to be useful to sophisticated peers.
Happy talk and showerthoughts were expressly verboten. As were any 101-level explainers. While presenters could certainly reference legitimate issues like hallucinations, IP, and privacy, they were limited to doing so in the context of decisions they actually made—e.g. we achieved acceptable accuracy by doing X…we did not extend the use case to Y due to privacy challenges we have yet to solve.
We were so strict on quality control we removed multiple law departments and law firms from the agenda when they could not secure the internal clearances to share at the required level of detail. Indeed, beyond welcome messages and housekeeping, LexFusion did not present at our own events because we do not meet the criteria—while we certainly use GenAI in many areas of our business, we do not have our own legal use cases (thankfully).
Our September New York event was covered in a wonderful two-part series (here and here) by Stephanie Wilkins, Editor in Chief of Legaltech News.
Dan Hoadley, Head of Data Science & Analytics at Mishcon de Reya, opened his presentation near the end of the day in New York with some extremely kind words he has permitted us to share, “I am so thankful to be here. I have learned more today than I have in the last decade of legaltech conferences, combined.” Dan then backed up those words by convincing us to put on the London version in November and convincing his own firm to host, which we ran alongside Dan Sinclair and the MDR Lab team.
In between New York and London, our good friend Léo Murgel, SVP & COO of Legal and Corporate Affairs at Salesforce, hosted in San Francisco. Mike Haven, AGC/Head of Legal Operations at Intel and President of CLOC, attended San Francisco and was generous enough to proclaim, “Best executive-level event of the year, by far. If LexFusion invites you to anything, go.”
Demand for real talk proved high. In fact, we so miscalculated the positive response rate for the first event in New York we had to shut down invites about a third of the way through our Tier 1 invite list—with over half of the 100+ attendees traveling in from out of state. While we grew more deliberate, both London and SF still ended up standing-room only, with waitlists.
Truth be told, the use cases themselves are pretty boring.
Indeed, worse than boring, the truth is depressing. Turns out, GenAI is not magic. GenAI represents a legitimate advancement. Ignoring GenAI would be foolish. Failure to start down the GenAI path is a terrible plan. But GenAI is a path, not a teleportation device. There is a staggering amount of work to be done to realize GenAI’s potential, especially in an enterprise environment.
Echoing the prevalent conference patter that long preceded the GenAI hype, the real people doing the real work emphasize the importance of prioritization, planning, scoping, knowledge management, change management, leadership buy-in, data hygiene, training, maturation of the application layer, UI/UX…that is, all the hard parts that have always been non-negotiable impediments to tech-centric improvement initiatives.
In short, the roundtable participants explained in a variety of ways that successful deployment of GenAI is dependent on successful projects. Worthwhile projects. Essential projects. But projects just the same.
Projects, however, are problematic. Only 0.5% of projects deliver on time, on budget, and with the expected benefits.* Projects are, in short, unreliable—a concept worth exploring in the context of strategy.
Strategery
"Courage is knowing it might hurt and doing it anyway. Stupidity is the same.
That’s why life is hard."
Jeremy Goldberg
Strategy is similar. Strategy is subject to a fundamental tension.
is the production of a consistent, replicable outcomes. Validity is the production of outcomes that achieve strategic objectives. Reliability is essential. But there is constant temptation to confuse reliability with validity due to our natural impulse to maintain control—setting targets we are confident we can hit. While circumstances may sometimes dictate we focus on reliability (e.g., cost cutting) at the expense of validity (e.g., long-term investment), if that is all we ever do, we are not playing to win—we are only playing to lose more slowly.
requires placing bets—trafficking in the probabilities, not certainties, that can deliver valid outcomes. While all bets entail risk, some bets are sounder than others. But all bets must contend with the frustrations intrinsic to our VUCA (volatile, uncertain, complex, ambiguous) operating environment, which confounds our capacity to accurately calculate our odds. Win probabilities must not only incorporate our internal ability to execute but also external factors we may be able to influence but can never control. Independent decisions by customers, competitors, and regulators complicates strategic choices, and make them all the more important.
We are inherently loss averse. Loss aversion activates skepticism. Unchecked, skepticism makes quick work of uncertainty. Wherever there is uncertainty, there is judgment. Judgment is, by definition, open to legitimate challenge. Since nothing truly innovative by can proven in advance analytically, the risks of being wrong are not only real but easily recognizable. We can always identify why something might not work in a manner that is accurate and, without attempting it, irrefutable. Loss aversion therefore biases us towards the status quo.
Being the default, the status quo is often exempted from the strict scrutiny to which change is subjected. Yet maintaining the status quo is as much a strategic choice as any other. It just does not feel that way because succumbing to inertia is less energy intensive. But risk-free, trade-off-free options do not exist and will not be surfaced by process of elimination. Meanwhile, short-term is easy often becomes long-term hard because strategic choices are circumscribed by positioning—every day, every development, and every decision further constrains the spectrum of strategic choice.
We believe the era of stability is over. Maintaining the status quo to the maximum possible degree is no longer the safest bet. Change is risky. But not changing is riskier. Generative AI is an accelerant of structural trends that were already laying bare the limits of the existing operating models.
Within this context, it merits revisiting the definition of a project*:
Projects involve a series of planned activities designed to generate a deliverable (a product, a service, an event). These activities—which can be anything from a grand strategic initiative to a small program of change—are limited in time. They have a clear start and end; they require an investment, in the form of capital and human resources; and they are designed to create predetermined forms of value, impact, and benefits. Every project has elements that are unique. That’s key: Each contains something that has not been done before.
The era of stability is ending. Unprecedent opportunities abound for those positioned and prepared to seize them. But hard choices will be even more abundant, and unavoidable—whether we are prepared or not.
Winning requires knowing you might be wrong and proceeding anyway. Losing is the same. That’s why strategy is hard.
